1.10 — When one bounded public-index detail candidate fails but a later candidate for the same query yields a validated live record, retain the raw diagnostic but classify the terminal result as COMPLETE rather than PARTIAL. Failures for unrelated or unrecovered queries remain visible. 1.9 — Start the same bounded public-index recovery when a native Goofish search navigation times out before it yields a Baxia payload; successful recovery resolves that native failure rather than marking an otherwise complete dataset partial. 1.8 — Reserve the final result slot before an asynchronous dataset/PPE write, preventing concurrent fallback details from exceeding maxResults or its event cap. 1.7 — Add a bounded Google SERP discovery recovery when native Goofish search is Baxia-blocked and the managed index source is unavailable or incomplete. Candidate detail records must still match the requested query before they are saved or charged. Search requests no longer repeat a known Baxia block. 1.6 — Disable the experimental UNBLOCKER fallback by default: live Camoufox canaries show a TLS certificate failure before the public page, so it must not add latency to customer runs. 1.5 — Mark the UNBLOCKER proxy route as HTTPS-intercepting so Crawlee passes its trusted-context options to Camoufox; the generic status probe did not set that route fact. 1.4 — Allow the Camoufox recovery context to accept Apify Unblocker's proxy certificate, fixing SEC_ERROR_UNKNOWN_ISSUER before the retry reaches the public page. 1.3 — Ensure the bounded Apify Unblocker recovery route receives the blocked public URLs rather than being deduplicated by Crawlee's shared request queue. 1.2 — Add one bounded Apify Unblocker retry after the sticky-residential Camoufox route is blocked by Goofish Baxia; record fallback telemetry and preserve honest zero-result failures.
